U.S. patent application number 12/557335 was filed with the patent office on 2010-03-25 for three-way handshake (3whs) optical network signaling protocol.
This patent application is currently assigned to TELCORDIA TECHNOLOGIES, INC.. Invention is credited to Arnold Neidhardt, Ronald A. Skoog, Brian Wilson.
Application Number | 20100074623 12/557335 |
Document ID | / |
Family ID | 42005470 |
Filed Date | 2010-03-25 |
United States Patent
Application |
20100074623 |
Kind Code |
A1 |
Skoog; Ronald A. ; et
al. |
March 25, 2010 |
THREE-WAY HANDSHAKE (3WHS) OPTICAL NETWORK SIGNALING PROTOCOL
Abstract
A method for optical network signaling processing of a signal
from a first node to an end node through intermediate nodes is
presented. The method comprises determining, in a first pass from
the first node to the end node, available wavelengths and
wavelength conversion at each node, the end node optimizing
wavelengths using the available wavelengths and wavelength
conversions, at each node, dropping a cross-connect command, in a
second pass from the end node to the first node, choosing
wavelengths for connection based on the optimizing step, in a third
pass from the first node to the end node, receiving at each node a
signal message and releasing unused cross-connect commands, the end
node identifying the chosen wavelengths and releasing the unused
resources, and transmitting the signal on the chosen wavelengths.
Restoration paths can also be determined. Optimizing can include
selecting and marking one or more backup wavelengths.
Inventors: |
Skoog; Ronald A.; (Bend,
OR) ; Neidhardt; Arnold; (Middletown, NJ) ;
Wilson; Brian; (Rumson, NJ) |
Correspondence
Address: |
TELCORDIA TECHNOLOGIES, INC.
ONE TELCORDIA DRIVE 5G116
PISCATAWAY
NJ
08854-4157
US
|
Assignee: |
TELCORDIA TECHNOLOGIES,
INC.
Piscataway
NJ
|
Family ID: |
42005470 |
Appl. No.: |
12/557335 |
Filed: |
September 10, 2009 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61095749 |
Sep 10, 2008 |
|
|
|
Current U.S.
Class: |
398/79 |
Current CPC
Class: |
H04J 14/0295 20130101;
H04Q 11/0062 20130101; H04J 14/0212 20130101; H04Q 2011/0088
20130101 |
Class at
Publication: |
398/79 |
International
Class: |
H04J 14/02 20060101
H04J014/02 |
Claims
1. A method for optical network signaling processing of a signal
from a first node to an end node through one or more intermediate
nodes, comprising steps of: determining, in a first pass from the
first node to the end node, available wavelengths and available
wavelength conversion at each node of the first node and the one or
more intermediate nodes; optimizing a selection of preferred and
backup paths, by the end node in the first pass from the first node
to the end node, in accordance with the available wavelengths and
wavelength conversions; dropping, in a second pass from the end
node to the first node, a cross-connect command at each node of the
one or more intermediate nodes; choosing, by the first node, in the
second pass from the end node to the first node, wavelengths for
connection based on the optimizing step; receiving, in a third pass
from the first node to the end node, a signal message and releasing
unused cross-connect commands at each node of the first node and
the one or more intermediate nodes; identifying by the end node, in
the third pass from the first node to the end node, the chosen
wavelengths and releasing the unused resources; and transmitting
the signal on the chosen wavelengths.
2. The method according to claim 1, the step of optimizing
comprising steps of: determining by the end node, in the first pass
from the end node to the first node, a number of wavelengths and
marking the determined wavelengths as preferred; and selecting and
marking by the end node, in the first pass from the first node to
the end node, one or more backup wavelengths.
3. The method according to claim 2, wherein the wavelengths chosen
are the determined wavelengths marked as preferred.
4. The method according to claim 1, further comprising a step of
determining, in the first pass from the first node to the end node,
restoration wavelengths.
5. The method according to claim 4, wherein the step of determining
further comprises steps of: establishing at the first node the
available wavelengths and available wavelength conversion and the
available restoration paths; getting and forwarding, at each node
of the one or more intermediate nodes, the available wavelengths
and the available wavelength conversion; and getting and
forwarding, at each node of the one or more intermediate nodes, the
available restoration paths.
6. The method according to claim 1, wherein the step of determining
further comprises steps of: establishing at the first node the
available wavelengths and available wavelength conversion; and
getting and forwarding, at each node of the one or more
intermediate nodes, the available wavelengths and the available
wavelength conversion.
7. The method according to claim 1, wherein the end node
initializes the second path and the first node initializes the
third pass.
8. A computer readable medium having computer readable program for
operating on a computer for optical network signaling processing of
a signal from a first node to an end node through one or more
intermediate nodes, said program comprising instructions that cause
the computer to perform steps of: determining, in a first pass from
the first node to the end node, available wavelengths and available
wavelength conversion at each node of the one or more intermediate
nodes; optimizing a selection of, by the end node in the first pass
from the first node to the end node, in accordance with the
available wavelengths and wavelength conversions; dropping, in a
second pass from the end node to the first node, a cross-connect
command at each node of the one or more intermediate nodes;
choosing, by the first node, in the second pass from the end node
to the first node, wavelengths for connection based on the
optimizing step; receiving, in a third pass from the first node to
the end node, a signal message and releasing unused cross-connect
commands at each node of the first node and the one or more
intermediate nodes; identifying by the end node, in the third pass
from the first node to the end node, the chosen wavelengths and
releasing the unused resources; and transmitting the signal on the
chosen wavelengths.
9. The program according to claim 6, the step of optimizing
comprising steps of: determining by the end node, in the first pass
from the end node to the first node, a number of wavelengths and
marking the determined wavelengths as preferred; and selecting and
marking by the end node, in the first pass from the first node to
the end node, one or more backup wavelengths.
10. The program according to claim 7, wherein the wavelengths
chosen are the determined wavelengths marked as preferred.
11. The program according to claim 6, further comprising a step of
determining, in the first pass from the first node to the end node,
restoration wavelengths.
12. The method according to claim 11, wherein the step of
determining further comprises steps of: establishing at the first
node the available wavelengths and available wavelength conversion
and the available restoration paths; getting and forwarding, at
each node of the one or more intermediate nodes, the available
wavelengths and the available wavelength conversion; and getting
and forwarding, at each node of the one or more intermediate nodes,
the available restoration paths.
13. The program according to claim 7, wherein the step of
determining further comprises steps of: establishing at the first
node the available wavelengths and available wavelength conversion;
and getting and forwarding, at each node of the one or more
intermediate nodes, the available wavelengths and the available
wavelength conversion.
14. The program according to claim 7, wherein the end node
initializes the second path and the first node initializes the
third pass.
Description
CROSS REFERENCE TO RELATED APPLICATIONS
[0001] The present invention claims the benefit of U.S. provisional
patent application 61/095,749 filed Sep. 10, 2008, the entire
contents and disclosure of which are incorporated herein by
reference as if fully set forth herein.
FIELD OF THE INVENTION
[0002] The present invention relates generally to dynamic optical
networks.
BACKGROUND OF THE INVENTION
[0003] A Wavelength Division Multiplex (WDM) optical network
consists of optical switches, fiber connecting the optical
switches, and WDM technology used to carry multiple wavelengths
(optical channels) in a fiber. The optical switches are either
Reconfigurable Optical Add-Drop Multiplexers (ROADMs) or Optical
Cross Connects (OXCs). ROADMs can be viewed as small OXCs, i.e.,
they connect to a small number of fibers. ROADMs and OXCs have
add/drop ports that connect to client ports, and optical
connections between client add/drop ports are set up through the
ROADM and OXC optical switching fabrics.
[0004] In WDM optical networks, an optical connection is set up
through multiple fibers. A basic connection consists of a single
wavelength channel, and it is desirable for the frequency of the
wavelength channel to be the same frequency in each fiber the
connection goes through. If a different frequency is used in
adjacent fibers, a wavelength converter is required, which
increases cost. The ROADMs and OXCs cross-connect the wavelength
used by the connection from one fiber to the other. In order for a
single wavelength to be used end-end for a connection, there must
be a fiber path between the connection end points that has that
wavelength available on each fiber in the path, i.e., it is not
being used for another connection on any of the fibers along the
fiber path. This is known as the "Wavelength Continuity Constraint"
(WCC).
[0005] If a single wavelength is not available in each fiber along
a fiber path, the connection can be established using wavelength
conversion connecting two fibers that require different wavelengths
within the ROADMs or OXCs. It is desirable to minimize the amount
of wavelength conversion required, since the wavelength conversion
is done with expensive opto-electronic equipment. Thus, an
important part of setting up optical connections is having
information available to be able to determine what wavelengths are
available in the different fibers and what OXCs and/or ROADMs have
available wavelength converters. With this information, end-end
optical connections can be established.
[0006] In addition to meeting the WCC and minimizing the amount of
wavelength conversion that is required, it is necessary to size the
network, e.g., number of wavelengths per fiber and number of fibers
between OXC and/or ROADMs, to meet blocking requirements for
on-demand services. Typical blocking requirements are on the order
of 10**-2 to 10**-3.
[0007] Another aspect of setting up optical connections is that
some services provide restoration after a failure; such as a fiber
cut, causes the working channel to fail.
[0008] Dynamic and/or WDM optical networks require optical
connections meeting the Wavelength Continuity Constraint (WCC) and
minimizing the amount of wavelength conversion required when the
WCC cannot be met. Further, setting up connections very quickly,
e.g., ranging from 100 ms to a few seconds, is desirable. Previous
methods to perform these tasks have been based on the work done by
the Internet Engineering Task Force (IETF) in their Generalized
Multi-Protocol Label Switching (GMPLS) standards, and in particular
RFC 3471, which gives the GMPLS signaling functional description,
and RFC 3473, which defines the Resource Reservation
Protocol-Traffic Engineering (RSVP-TB) signaling procedures.
[0009] Probing techniques have been used to collect recent
information on available resources. However, there is a relatively
high likelihood that the resources identified by the probes as
being available may actually not be available when the reservation
request arrives. GMPLS-like methods need to do distributed
processing with the RESERVE message, which means processing at a
number of nodes along the connection setup path.
[0010] The prior art (GMPLS) was fixated on more distributed
processing methods between Node A and Node Z, where information is
passed, usually in a PATH message, from one node to the next, and
processing is done along the way so that when the PATH message
reaches Node Z, the wavelength to use to the first upstream node
can be determined. Subsequent decisions of wavelengths and
wavelength conversion are then made at each node along the RESERVE
message path from Node Z to Node A. This distributed processing
paradigm results in slow connection setup times, and sub-optimal
decisions are made. However, it has been the preferred paradigm to
use, and this has been strongly influenced by the work in the
IETF.
[0011] GMPLS methods probe only a single working path, which is
usually determined by link-state update information. Link-state
update information occurs on a relatively slow time scale, so there
is a reasonable probability there are other paths that are more
optimal. The GMPLS methods choose working and restoration paths
based on slower link-state advertisements that provide summary
information on link state.
[0012] One of the means that has been considered to distribute
information about available wavelength on fiber links is using the
IETF extensions to (Open Shortest-Path First Interior Gateway
Protocol) OSPF to support GMPLS, which are provided in RFC 4203.
The difficulty with these techniques is that the OSPF link state
updates cannot be sent out too frequently, so the information
becomes stale very quickly in very dynamic networks. A method to
get more up-to-date link state information for wavelength services
is defined in RFC 3473 which defines a Label Set Object. This Label
Set Object collects more current information than GMPLS link state
updates can provide, but in dynamic networks its information is
still somewhat stale.
[0013] There are three major deficiencies in previous methods. One
deficiency is that the information the methods use, e.g., the Label
Set Object, to collect available resource infatuation, for example
available wavelengths in each fiber, leads to stale and incomplete
information at the end node that uses the information to choose the
wavelength to use for the connection. As a result, there is a
non-negligible probability the chosen wavelengths will not be
available along the entire selected path when the reservation step
of the procedure is executed. Information is incomplete because the
Label Set Object only provides Node Z with available wavelength
information beyond the last node on the path that does wavelength
conversion, and it provides no information on available wavelength
converters at the nodes along the signaled path.
[0014] The second major deficiency in previous methods is that they
do a poor job of optimizing, e.g., minimizing, the number of
wavelength converters required to set up a connection. One proposed
improvement is called a "Suggested Vector" which does provide
significant improvement in the consideration of wavelength
conversion. However, the Suggested Vector does not consider the
number of available wavelength converters in individual nodes,
limiting its optimization capabilities.
[0015] The third major deficiency is the inability of the previous
methods to set up connections in a very short amount of time, e.g.,
within 100 ms in the Continental US. The previous methods, GMPLS
and extensions based on GMPLS, need to store information like label
sets, suggested labels, suggested vectors, etc., in each node along
the path. This is because the end node (Z end) does not have enough
information to determine what wavelengths are to be used on each
link and where wavelength conversion is to be done. The Z end just
knows which wavelength to use on the hop to the first up-stream
node. The subsequent wavelength and wavelength conversion decisions
at the upstream nodes are determined from the previously stored
information from the downstream (A to Z) PATH message. This means
that when the RESERVE message goes from Node Z to Node A to set up
the connection, there must be logical processing to determine what
needs to be done at each node. This requires significant processing
time, and thus results in relatively slow connection setup
times.
[0016] A procedure is needed that does the logical processing only
once (at Node Z), and on the signaling pass from Node Z to Node A
(Pass 2) very simple cross-connect and wavelength conversion
commands can be given to the switches, enabling very fast signaling
propagation times between Nodes A and Z.
SUMMARY OF THE INVENTION
[0017] An inventive method for signaling protocol that can
compensate for stale information, set up optical connections very
quickly, and achieve low blocking probability and efficient
resource (wavelengths, wavelength converters) usage is presented.
The inventive procedure is advantageous in terms of the amount of
capacity (wavelengths per fiber) that is needed to meet blocking
requirements. Also, the inventive procedure allows for much faster
signaling speeds, and it enables the optimization of the use of
wavelength converters. Further, extra or backup connections are
reserved, which significantly reduces the probability of blockage
with a negligible impact on increased resource usage.
[0018] The inventive system and method allows a consideration of
the number of available wavelength converters in the nodes. That
is, the consideration of where to do wavelength conversion to
include the current availability of wavelength converters along the
connection path is enabled. This can have a significant effect on
blocking performance. For example, if one node has very few
wavelength converters left and another has many more available, and
if the connection being set up can be made by doing wavelength
conversion at either node, it is highly desirable to do the
wavelength conversion at the node having the larger number of
converters. This leaves more wavelength converters at the node with
only a few left, and thus it reduces the likelihood that that node
will run out and subsequently cause blocked calls due to the
inability to do wavelength conversion at that node. The invention
can make these kinds of assessments, and previous procedures
cannot.
[0019] The inventive system and method can also include a type of
restoration, called "Shared Mesh Restoration", in which a
restoration path that is diverse from the working path is
determined as part of the connection provisioning process. The
restoration paths are only set up after a failure occurs, so if two
working connections do not share any failure nodes, they can both
"share" the same restoration resources. Thus, for provisioning
connections using shared mesh restoration, it is important to be
able to identify what wavelengths on different fibers can be shared
for restoration.
[0020] The inventive method does simple gathering of information
from the nodes and does all the significant logical processing at
the end nodes (Nodes Z and A). In the past such solutions were
avoided because of the limited processing capabilities available,
but today there is significant processing available on very small
chips, and the arguments for the more distributed processing
approach are no longer valid.
BRIEF DESCRIPTION OF THE DRAWINGS
[0021] The invention is further described in the detailed
description that follows, by reference to the noted drawings by way
of non-limiting illustrative embodiments of the invention, in which
like reference numerals represent similar parts throughout the
drawings. As should be understood, however, the invention is not
limited to the precise arrangements and instrumentalities shown. In
the drawings:
[0022] FIG. 1 illustrates the flow of the inventive method;
[0023] FIG. 2 is a flow diagram of a first path in a first
embodiment;
[0024] FIG. 3 is a flow diagram of a first path in a second
embodiment;
[0025] FIG. 4 is a flow diagram of a second path in the first
embodiment;
[0026] FIG. 5 is a flow diagram of a second path in the second
embodiment; and
[0027] FIG. 6 is a flow diagram of a third path of the inventive
method.
DETAILED DESCRIPTION
[0028] An inventive method for signaling protocol that can
compensate for stale information, set up optical connections very
quickly, and achieve low blocking probability by having alternate
paths, and efficient resource, e.g., wavelengths, wavelength
converters, usage by holding paths for the shortest time and
minimizing use of wavelength converters is presented. The signaling
protocol process of the present invention is one which probes for
available wavelengths as part of the circuit setup process. This
signaling protocol efficiently gathers the information needed to
set up working and shared mesh restoration paths in a manner that
allows very fast circuit setup times, e.g., 100 ms.
[0029] As an example case, assume that connections requiring 1, 2,
4 or 8 wavelength channels can be requested. The paths to probe can
be determined from GMPLS link state information providing aggregate
available capacity on the fiber links. The aggregate spare capacity
changes very slowly in accordance with the changes in aggregate
traffic intensity, so this link state information becomes stale on
a much longer timescale, e.g., minutes, and therefore requires much
less frequent updates.
[0030] The concept of the inventive signaling protocol is
illustrated in FIG. 1 for a single fiber path. Note that multiple
fiber paths (not shown) could be probed simultaneously. The first
signaling pass P1 (Node A to Node Z) collects data (X data and
wavelength converter data) from each optical switch, e.g., OXC,
OADM, along the fiber path, giving the available wavelengths in
each fiber pair and the available wavelength conversion resources.
When the signaling message arrives at Node Z, that end node can
determine very quickly which wavelengths are available along the
entire fiber path. Z then picks the number of wavelengths required
for the connection (1, 2, 4, or 8) from the available wavelengths,
and identifies these as preferred. It also picks a small number of
additional (backup) wavelengths, discussed below, from the
available wavelengths to allow for contention with other
connections to occur without those contentions causing
blocking.
[0031] Node Z then initiates its local cross-connect commands or
cross-connects, described below, from the add/drop ports to the
selected wavelengths, and also initiates the second signaling pass
P2 (Z to A). The wavelengths chosen would be in accordance with a
wavelength assignment (WA) strategy such as First Fit. If a
continuous wavelength is not available, wavelength conversion will
be used; Node Z will identify the wavelength conversion nodes and
the wavelengths to use in the all-optical segments. Note that the
number of backup wavelengths might be one or two for single
wavelength connections, and for multi-wavelength connections, it
could be more.
[0032] In the second signaling pass P2 (Z to A) the signaling
message drops cross-connects, and wavelength conversion information
if needed, at each intermediate node to cross-connect the selected
wavelengths, both the preferred and backup. The signaling message
does not wait for the cross-connect to complete, it keeps going.
The signaling node does check if the required wavelengths are still
available and marks connections that try to use unavailable
resources as failed. When the signaling message arrives at Node A,
it chooses a set of the successful wavelengths for the connection.
Note that if the preferred wavelengths are successful, then they
are chosen.
[0033] Node A initiates the third signaling pass P3 back to Z, and
Node A initiates its cross-connect from the add/drop ports to the
chosen wavelengths. When those connections to the add/drop ports at
Node A are complete, the client at Node A can begin to send
data.
[0034] In the third signaling pass P3 (A to Z), when the signaling
message is received by the intermediate nodes, the cross-connects
for the unused wavelengths are released. When the third pass
signaling message arrives at Node Z, it identifies the chosen
wavelengths and releases the unused wavelengths.
[0035] The collection of link state information in the first
signaling pass is done using a very fast write by the node element
(NE) into the signaling packet, e.g., a vector of 0s and 1s could
be written for each fiber-pair into the signaling message; with 0
representing wavelength unavailability on that fiber-pair. The
signaling packets might travel on a signaling network that uses a
dedicated signaling Optical Supervisory Channel (OSC) in each fiber
and a dedicated signaling packet switch capability in each NE.
Thus, this write process can be done very fast at near hardware
speeds (e.g., <0.1 ms).
[0036] A cross-connect is an optical switch configuration where a
signal from a specific wavelength on one port, e.g., the A end, is
delivered to a specific wavelength on another port, e.g., the Z
end, possibly using wavelength conversion resources. A
cross-connect command tells the optical switch to setup one or more
cross-connects. That is, some switches may allow multiple
cross-connects to be configured with a single command while others
may not. The present invention is not limited to this cross-connect
which is described merely for exemplary purposes.
[0037] When shared mesh restoration is used, the inventive
procedure is enhanced to also probe the candidate restoration paths
to then select restoration as well as working paths and
wavelengths. The basic model is that wavelengths in a fiber link
are in one of the following states: unreserved, e.g., idle,
in-service, or shared-reserved, i.e., shared by multiple
restoration paths. It can be assumed that there is a path
computation process that determines candidate pairs of working and
restoration paths. That is, working paths are computed for each A-Z
node pair, and for each working path there are one or more
candidate restoration paths computed. The restoration paths must be
disjoint from the working path, but the candidate restoration paths
for a particular working path do not need to be disjoint.
[0038] When a connection request arrives, the inventive method is
used to probe with Pass 1 messages one or more candidate working
paths for available (unreserved) wavelengths and wavelength
converters. At the same time, the restoration paths are probed to
identify wavelengths in the unreserved and shared reserved state
that can be used to protect the working path. Also Pass 1 P1 probes
collect information on wavelength converters along the restoration
paths. Since the working and restoration paths are pre-computed,
the nodes along the restoration paths can be provisioned with the
Shared Risk Link Groups (SRLGs) of each working path the node is
supporting. An SRLG identifies an entity that can fail, such as a
fiber, node, conduit, tunnel, bridge, etc. When a restoration path
is established, the wavelengths in the restoration path are put in
the shared reserve state and the SRLGs the restoration path is
protecting are stored locally.
[0039] When a Pass 1 message goes along a restoration path, the
message picks up those wavelengths in the shared reserve state that
are not currently protecting the SRLGs of the working path. Note,
as described above, the working path SRLGs are available locally
along the restoration path. The Pass 1 message also picks up the
wavelengths in the unreserved state and wavelength converter
availability information. On the working paths, the Pass 1 messages
operate as described above. When the working and backup path Pass 1
messages all arrive at Node Z, an algorithm is run to select the
best working/backup pair for the connection request. For the choice
of restoration paths, the main objective is to maximize sharing, so
it is desired to maximize the number of wavelengths in the shared
reserved state. Other metrics can also be used, such as number of
working paths being protected by wavelengths in the shared reserve
state. The use of wavelength converters is also optimized.
[0040] On Pass 2 P2, extra resources can be reserved on both the
working and backup paths to reduce the probability of blocking from
contention with other connections being set up. Node A would make
the final selection of working/restoration paths and Pass 3
messages would release the unused Pass 2 P2 reservations.
[0041] An advantage of this inventive procedure is that on Pass 2
P2 (Node Z to Node A), extra channels are reserved to protect
against getting blocked from resources, identified as idle when
Node Z made its selection, becoming busy. Analysis using a
requirement of 10**-3 blocking probability shows that the previous
procedures without reserving extra resources on Pass 2 P2 would
require approximately five times as many wavelengths in each fiber
as the inventive method requires. Moreover, reserving extra
resources on Pass 2 P2 increases resource usage by only about one
percent.
[0042] As discussed above, the inventive process does very simple
data collection on Pass 1 P1, and very simple commands on Pass 2 P2
and Pass 3 P3. All of the logical processing affecting setup time
is done once at Node Z after Pass 1 P1. As a result, very fast
connection setup times are possible. Previous methods do logical
processing at each switch for both Pass 1 (the A to Z PATH message)
and Pass 2 (the Z to A RESERVE message). This results in much
slower signaling propagation times.
[0043] In the setup of shared mesh restoration paths, the inventive
procedure does not require collecting SRLG information in the
signaling probes. Due to pre-provisioning the working path SRLGs in
the nodes of the working path's restoration paths, the Pass 1
probes on restoration paths only pick up shared reserve wavelengths
that do not protect SRLGs of the working path. Previous procedures
have to collect SRLG information and process that information at
Node Z. Hence the inventive methodology enables signaling
processing to be much simpler and quicker than that done with other
methods.
[0044] The extra channels reserved on Pass 2 P2 make a significant
difference (reduction) in the blocking probability without
consuming significant excess resource usage. The concept of, on
Pass 2, reserving more resources than needed for the requested
connection is a strategy that has a very high payoff without a
significant cost penalty in increased resource usage since these
resources are only reserved for a short time (a few
milliseconds).
[0045] FIG. 2 is a flow diagram of the first pass of the inventive
method. The pass P1 begins at step S1 with Node A. At step S1, Node
A determines working path pairs, and launches the pass 1 messages.
At step S2, for nodes between A-Z, e.g., intermediate nodes, at
each optical switch or node, the availability of wavelength and
wavelength conversion resources are obtained, that is, the working
path pairs from Node A are modified to incorporate the availability
of resources. Accordingly, each intermediate node forwards just one
message for each working path or candidate pair that passes through
that node. In other words, step S2 is performed once for Node B
(first node connected to node A), and for the next connected node,
and for all nodes until Node Z. Note that each working path being
explored by Node A has one or more intermediate nodes, with
relevant intermediate nodes being those on the working path(s). At
Node Z, optimization is performed in step S3, that is, the
wavelengths to use in Pass 2 P2 for working path pairs are
determined and marked as preferred. In step S4, Node Z initiates
pass 2 P2.
[0046] FIG. 3 is flow diagram of the first pass of the inventive
method in accordance with an embodiment incorporating restoration
paths in the inventive procedure. In Step S5, Node A determines
both working and restoration path pairs, and launches the pass 1
messages. At step S6, for nodes between A-Z, at each node, the
availability of wavelength and wavelength conversion resources for
working path and for restoration path are obtained and forwarded.
Each intermediate node forwards just one message for each path
pair, e.g., working and/or restoration, that passes through the
node. At step S7, Node Z determines which wavelengths to use in
Pass 2 for working and for restoration, and these wavelengths are
marked as preferred. Step S4 initiates Pass 2 P2 as in FIG. 2.
[0047] FIG. 4 is a flow diagram of the second pass P2. The pass P2
begins at step S8 as follows. At step 58, for nodes Z-A, at each
node, a drop of the cross-connect command is performed and extra
resources are reserved. At step S9, at Node A, the set of
successful wavelengths for connection are chosen. In step S 10,
Node A initiates pass 3 P3.
[0048] FIG. 5 is a flow diagram of the second pass P2 in the
embodiment including restoration paths. In this embodiment, both
the steps shown in FIG. 4 and those shown in FIG. 5 are performed.
The pass P2 begins at step S11 as follows, and can be performed
before, after, or in conjunction with steps S8 and S9. At step S11,
for, nodes Z-A, at each node, a drop of the restoration
reservations is performed and extra resources are reserved. At step
S12, at Node A, the set of successful wavelengths for connection
are chosen. In step S13, Node A initiates pass 3 P3.
[0049] FIG. 6 is a flow diagram of the third pass P3. The pass P3
begins at step S14 with Node A. At step S14, for nodes A-Z, at each
node, the signal message is received and unused cross-connect
resources are released. At step S15, at Node Z, the chosen
wavelength is identified and any remaining unused resources are
released. At step S16, the signal is transmitted along the chosen
wavelengths and wavelength conversions.
[0050] An example of the inventive procedure follows. One measure
of the efficiency of a signaling and wavelength selection protocol
is the number of wavelengths required on the fiber links to achieve
a specified blocking probability. Given a 3-hop fiber path between
Nodes A and Z, and an aggregate load (from all paths) on each fiber
link to be a 10 Erlang load. Consider how many wavelengths would be
required on a fiber to achieve a 10-3 blocking on the 3-hop path.
As a baseline for comparison, use the perfect case of zero
propagation and processing delays and complete information on
available wavelengths. In that case, 28 wavelengths are required to
achieve 10-3 blocking probability.
[0051] If the round-trip propagation time were to be 0.01 times the
mean call holding time, then if no extra wavelengths are reserved
on Pass 2 P2, 161 wavelengths are required to achieve 10-3 blocking
probability. However, if a single extra wavelength is reserved on
Pass 2 P2, 28 wavelengths are required, just as in the ideal, zero
propagation delay, case. With the extra wavelength being reserved
on Pass 2 P2, the average link load becomes 10.1. Erlang, so the
inventive procedure increases the load by just one percent.
[0052] Other experimental results similarly show that the inventive
method needs just one or two extra wavelengths on Pass 2 to achieve
the same performance as the ideal, zero propagation delay, case. We
also note that RSVP-TE signaling techniques using the IETF
standardized Shared Label set to identify available wavelengths
would require the 161 wavelengths rather than the 28 required by
the present invention.
[0053] A simulation study was done to compare the inventive
procedure with the standardized GMPLS signaling methods in terms of
the use of wavelength converters. It was shown that the GMPLS
methods require 74% more wavelength converters. Further, the
present invention advantageously considers the number of available
wavelength converters in individual nodes, whereas prior
techniques, including the "suggested vector" technique discussed
above, do not.
[0054] Accordingly, advantages of the present invention include the
following. A powerful new signaling protocol procedure for dynamic
optical networks is provided. The procedure enables very fast setup
time, low backward blocking and efficient restoration. The
inventive computation model is not distributed, instead it collects
all required information and then computes optimal solution in one
operation. The inventive technique has explicit (adaptive) control
over backwards blocking by the selection of the number of extra
channels that are reserved on Pass 2 P2. Multiple paths are probed,
enabling the best working path to be chosen based on current,
detailed information. Based on current network state, both working
and restoration paths can be chosen at the same time. In addition,
all optical segments can be maximized and wavelength conversions
can be minimized.
[0055] Various aspects of the present disclosure may be embodied as
a program, software, or computer instructions embodied in a
computer or machine usable or readable medium, which causes the
computer or machine to perform the steps of the method when
executed on the computer, processor, and/or machine A program
storage device readable by a machine, tangibly embodying a program
of instructions executable by the machine to perform various
functionalities and methods described in the present disclosure is
also provided.
[0056] The system and method of the present disclosure may be
implemented and run on a general-purpose computer or
special-purpose computer system. The computer system may be any
type of known or will be known systems and may typically include a
processor, memory device, a storage device, input/output devices,
internal buses, and/or a communications interface for communicating
with other computer systems in conjunction with communication
hardware and software, etc.
[0057] The terms "computer system" and "computer network" as may be
used in the present application may include a variety of
combinations of fixed and/or portable computer hardware, software,
peripherals, and storage devices. The computer system may include a
plurality of individual components that are networked or otherwise
linked to perform collaboratively, or may include one or more
stand-alone components. The hardware and software components of the
computer system of the present application may include and may be
included within fixed and portable devices such as desktop, laptop,
and server. A module may be a component of a device, software,
program, or system that implements some "functionality", which can
be embodied as software, hardware, firmware, electronic circuitry,
or etc.
[0058] The embodiments described above are illustrative examples
and it should not be construed that the present invention is
limited to these particular embodiments. Thus, various changes and
modifications may be effected by one skilled in the art without
departing from the spirit or scope of the invention as defined in
the appended claims.
* * * * *